Advancing AI Workflows and Inference with DevOps, MLOps, OSCAR and AI4Compose

Friday Apr 11, 2025, 10:00 AM → 12:00 PM Europe/Madrid

Judith Sainz-Pardo Diaz, Marcin Plociennik (PSNC), Álvaro López García (IFCA)

In this webinar, framed within the scope of the AI4EOSC project, we will explore tools and methodologies for optimizing AI development and deployment. First, we will give an introduction to DevOps practices in the software development life cycle. Next, we will analyze the MLOps principles, followed by a live demonstration on the use of MLFlow within the AI4EOSC platform. Participants will also learn about the OSCAR system for inference, and AI4Compose, a tool for seamlessly composing AI inference pipelines.

Through these talks and demos, the webinar highlights key tools for advancing and optimizing AI operations and workflows.

Streaming via YouTube: https://www.youtube.com/watch?v=t_qDCYQ31RI

10:00 AM → 10:30 AM Software development life cycle: introduction to DevOps

Speaker: Andres Heredia Canales (IFCA)

10:30 AM → 11:00 AM Introduction to MLOps and MLFlow demonstration

Speaker: Lisana Berberi (KIT)

11:00 AM → 11:30 AM OSCAR system for inference

Speaker: Amanda Calatrava Arroyo (Universitat Politècnica de València)

11:30 AM → 12:00 PM AI4Compose: composition of AI inference pipelines

Speaker: Diego Aguirre Ramirez (Universitat Politècnica de València)